BACKGROUND Although it has been used as a diagnostic tool in many renal transplant centres the reproducibility of fine-needle aspiration biopsy (FNAB) has not been critically evaluated. METHODS In the present study material sequentially obtained from 15 patients (177 aspirations) over a 3-month period following renal transplantation was evaluated by two independent observers. Intraobserver reproducibility was studied through the analysis of two evaluations by observer 1 performed 8 months apart. Interobserver reproducibility was calculated comparing the second evaluation of observer 1 with the evaluation of observer 2. All evaluations were performed blindly. Slides and protocols of individual patients were chronologically arranged and thus interpreted. Representation, total corrected increment (TCI), accuracy, and the percent agreement for the diagnosis of acute rejection were evaluated. RESULTS No intraobserver statistically significant differences were observed. Differences observed in interobserver evaluation of TCI either during, or out of acute rejection episodes, well as the representative quality of the sample and accuracy, were not statistically significant. The percent of agreement for the presence or absence of acute rejection was 77.5% for the intraobserver comparison and 76.7% for the interobserver evaluation. Both values showing either a non-significant value for the difference or a statistical significance for the concordance. CONCLUSIONS We concluded that FNAB is an accurate method with fairly good intra- and interobserver reproducibility for the diagnosis of acute rejection of renal allografts.